The immobilization of enzymes on atomic force microscope tip (AFM tip) surface is a crucial step in thedevelopment of nanobiosensors to be used in detection process. In this work, an atomistic modeling ofthe attachment of the acetyl coenzyme A carboxylase (ACC enzyme) on a functionalized AFM tip surface isproposed. Using electrostatic considerations, suitable enzyme–surface orientations with the active sitesof the ACC enzyme available for interactions with bulk molecules were found. A 50 ns molecular dynamicstrajectory in aqueous solution was obtained and surface contact area, hydrogen bonding and proteinstability were analyzed. The enzyme–surface model proposed here with minor adjustment can be appliedto study antigen–antibody interactions as well as enzyme immobilization on silica for chromatographyapplications.

1. Journal of Molecular Graphics and Modelling 45 (2013) 128–136
Contents lists available at ScienceDirect
Journal of Molecular Graphics and Modelling
journal homepage: www.elsevier.com/locate/JMGM
Molecular modeling of enzyme attachment on AFM probes
Guedmiller S. Oliveiraa,∗
, Fabio L. Leiteb
, Adriano M. Amaranteb
, Eduardo F. Francac
,
Richard A. Cunhac
, James M. Briggsd
, Luiz C.G. Freitasa
a
Chemistry Department, Federal University of São Carlos, 13565-905 São Carlos, SP, Brazil
b
Nanoneurobiophysics Research Group, Department of Physics, Chemistry and Mathematics, Federal University of São Carlos, 18052-780 Sorocaba, SP,
Brazil
c
Chemistry Institute, Federal University of Uberlândia, 38400-902 Uberlândia, MG, Brazil
d
Department of Biology and Biochemistry, University of Houston, 77204-5001 Houston, TX, USA
a r t i c l e i n f o
Article history:
Received 30 May 2013
Received in revised form 18 July 2013
Accepted 6 August 2013
Available online 16 August 2013
Keywords:
Immobilization
Atomic force microscope tip
Enzyme
Computational simulation
Model
Force field parameters
a b s t r a c t
The immobilization of enzymes on atomic force microscope tip (AFM tip) surface is a crucial step in the
development of nanobiosensors to be used in detection process. In this work, an atomistic modeling of
the attachment of the acetyl coenzyme A carboxylase (ACC enzyme) on a functionalized AFM tip surface is
proposed. Using electrostatic considerations, suitable enzyme–surface orientations with the active sites
of the ACC enzyme available for interactions with bulk molecules were found. A 50 ns molecular dynamics
trajectory in aqueous solution was obtained and surface contact area, hydrogen bonding and protein
stability were analyzed. The enzyme–surface model proposed here with minor adjustment can be applied
to study antigen–antibody interactions as well as enzyme immobilization on silica for chromatography
applications.
© 2013 Elsevier Inc. All rights reserved.
1. Introduction
The design of functionalized atomic force microscope tips
(AFM tips) to build biosensors with high definition and accuracy
requires the quantification of the interaction energetics between
organic and inorganic compounds, such as protein–surface and
protein–ligand complexes. The complexity of the adsorption
process of enzymes onto inorganic surfaces can be better under-
stood by a combination of theoretical and experimental studies.
Therefore, the achievements presented here are useful to build
biosensors capable to detect inhibitors (e.g. herbicides) by using
enzymes as a probe.
In a previous paper of our research group [1] we have been
developing prototypes of nanobiosensors to detect herbicides used
in agriculture. In addition, our obtained results from combining
theoretical and experimental data to investigate the interaction
of two pesticides (diclofop and atrazine) with the active sites of
the acetyl-coenzyme A carboxylase (ACC enzyme) were in a good
agreement [2–4]. In the present paper, the earlier studies were reas-
sumed and the aim is the development of a molecular model and
a computational protocol to study enzyme–surface interaction to
∗ Corresponding author. Tel.: +55 34 9174 2501.
E-mail address: guedmuller@gmail.com (G.S. Oliveira).
build biosensors. These investigations are part of a general research
project aimed to investigate natural products capable of interacting
with important enzymes of plants and animals to be efficiently used
in insect plague control [5].
AFM experiments have been successfully employed to measure
intra-molecular unfolding forces of individual proteins [6,7] and
intermolecular forces between various ligands and receptor pairs
[8–14]. Compared with other biophysical approaches for direct
measurements of intermolecular forces, such as optical tweezers,
magnetic torsion devices, and bio-membrane force probes (BFP),
AFM has the unique advantage of possessing the high spatial reso-
lution needed to detect molecular interaction. Additionally, recent
progress in AFM studies has made it possible to directly quantify the
range and magnitude of the interactions forces between proteins
and other molecules [15]. Nevertheless, the AFM technique is very
useful for quantifying single molecule interaction, but it does not
reveal neither molecular mechanisms behind the binding of ligands
nor conformational changes in biomolecules in atomic time scale
[1]. The goal of quantifying intermolecular forces, binding energies
as well as complex stabilization of protein–protein, protein–ligand
and protein–surface systems can be obtained by combining AFM
results with theoretical studies. Generally, the specificity of an
AFM tip is obtained by using enzymes as molecular probe. There-
fore, the combination of computational methods such as quantum
mechanics (QM) and molecular dynamics (MD) calculations with
1093-3263/$ – see front matter © 2013 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jmgm.2013.08.007

2. G.S. Oliveira et al. / Journal of Molecular Graphics and Modelling 45 (2013) 128–136 129
Fig. 1. Connection of the linker SSA with atoms from the AFM tip surface. The silicon and oxygen atoms (Si-O-Si) were position restrained. (For interpretation of the references
to color in this figure legend, the reader is referred to the web version of this article.)
experimental analysis will be useful to make optimal estimative of
enzyme immobilization to propose AFM tips with enhanced sen-
sibility. In this paper, we report the use of computer simulation
methodologies (MD and QM) to investigate proper conditions to
immobilize an enzyme as to build functionalized AFM tips.
The target enzyme used in this investigation was the ACC which
has crucial roles in fatty acid metabolism of humans and most
other living organisms. ACC is also an attractive target for drug
discovery against a variety of human diseases, including diabetes,
obesity, cancer, and microbial infections. In addition, ACC from
plants has been a target of herbicides used commercially for more
than 20 years. Haloxyfop, diclofop (FOPs), sethoxydim and butroxy-
dim (DIMs) are herbicides capable to occupy the carboxytransferase
(CT) domain of the ACC and hamper its enzymatic function.
AFM tips are made of silicon (Si) or silicon nitride (Si3N4) and,
in order to attach enzymes acting as a probe, this surface must be
modified. Among the existing methodologies to chemical modifi-
cation of AFM tip surface, two techniques have been used mostly
in the last few years. The first uses tips covered with gold [16] and
thiol entities which are properly bonded to biomolecules. The sec-
ond uses the oxidation of the AFM tip surface due to its exposure to
air or water, as a result a large amount of reactive Si OH groups are
formed reacting strongly with silanes, alcohols and also polymers
such as polyethylene glycol (PEG) [17,18]. In the present study, a
surface spacer agent (SSA) obtained by combining 3-(aminopropyl)
triethoxysilane (APTES) and glutaraldehyde [19,20] was used to
cover the AFM tip surface. For biosensor purposes, the glutaralde-
hyde reacts with APTES establishing a bridge for attachment of
biomolecules. Glutaraldehyde is a common cross-linking agent due
its reaction with residues such as lysine to form stable adducts
[20,21]. In order to ensure that the cross-linker molecules can effi-
cient adsorb an enzyme, the interaction forces between AFM tips
surface and enzyme must be greater than those between inhibitors
and proteins; otherwise, the enzyme might be pulled off from the
AFM tip. The linker provides stable adsorption conditions and flex-
ibility for biomolecules; as a result receptor–ligand interactions
capabilities are preserved. The geometric dimensions of the system
were chosen as to provide the appropriated computer simulation
conditions discussed below.
2. Methodology
In this work an atomistic model was proposed to study the inter-
action of the ACC enzyme with an AFM tip surface using molecular
dynamics methodology. In the next sections this atomistic model
and the procedures used to determine the force field (FF) parame-
ters are presented.
2.1. System definition
According to the experimental data reported by Etchegaray et al.
[19], Bhushan et al. [20], and Deda et al. [22], AFM tip probes can be
modified using APTES as a linker. When AFM tips are exposure to
environment conditions an oxidation occurs naturally; as a result,
the AFM surface is hydroxylated forming SiOH groups at the
top of the surface allowing the interaction with APTES molecules.
Further, this surface is treated with glutaraldehyde forming an
APTES–glutaraldehyde complex which is capable of reacting with
amino acids groups such as lysine and arginine. As can be seen in
Figs. 1 and 2, the linker is bounded to the SiOH surface considering
Fig. 2. Final structure of the SSA linker. The atoms labels were used in the molecular
geometry scanning and electrostatic potential calculations.

3. 130 G.S. Oliveira et al. / Journal of Molecular Graphics and Modelling 45 (2013) 128–136
Fig. 3. Description of the simulated system using molecular dynamics methodology.
System dimensions are shown for each component of the system.
the silicon atom of the surface. This model was called surface-
spacer agent (SSA). The interaction of this surface model with
biomolecules can be straightforwardly obtained using FF param-
eters allowing a computational insight in the physical–chemical
process occurring on AFM tips. In the next section the FF parameters
were calculated and this atomistic model is discussed.
2.2. Description of the surface model
The silanol (SiOH) surface representing the oxidized AFM tip
was modeled using a SiO2 network with hydrogen atoms bonded
to the oxygen of the top of the surface. The initial SiO2 geometry was
obtained using standard procedures available in Inorganic Builder
Plugin implemented in VMD program [23]. Schematic representa-
tions of the linker surface spacer agent (SSA) model on the surface
are shown in Figs. 1 and 2. The surface geometry was optimized
and positions obtained for some silicon atoms were used to dis-
tribute196 linker molecules on a squared surface with dimensions
194 ˚A × 194 ˚A as can be seen in Fig. 3. The silicon atoms bounded
to the linkers were constrained to their original surface position
during the simulations in order to represent solid state like strong
interactions. Table 1 summarizes the number of atoms of the sim-
ulated system.
The remaining atoms on the SiOH surface network were not
explicitly included in the simulation but their influence were con-
sidered using a potential function facility available in the NAMD
program [24], as discussed by Aksimentiev et al. [25]. This atom
less representation reduces the degrees of freedom of the system
decreasing the computational demands of the simulations.
Table 1
Number of particles (N) of each system component.
Component N
SSA model 6076
ACC enzyme 23,150
Tip3p water 143,198
Na+
contra-ions 30
Total 172,454
Table 2
Bond distances and bond angles (r0
Å , Â0
) and force constants (ks
, kb
) for the SSA
model reported in Fig. 2.
Bond (r) ks
(kcal mol
−1
Å
2
) r0
(Å
2
)
C8 Si 306.432 1.5080
Si O Fixed 1.9780
O Si Fixed 1.4105
Angle (Â) kb
(kcal mol
−1
rad
−2
) Â0
(◦
)
C8 Si O 418.400 113.050
C7 C8 Si 488.273 112.700
Si O Si Fixed 158.070
2.3. Surface-spacer agent (SSA) parameterization
To perform molecular dynamics simulations using a classical
FF approach to be implemented in OPLS-AA FF protocol [26], the
total energy is divided into two terms, the Eintra and Einter. The Eintra
interaction energy term is defined as:
Eintra
= Estrech
+ Ebend
+ ERots
+ EFtors
+ Enb
(1)
The first three terms have a harmonic expression:
Estrech
=
1
2
NS
Ks
(r − r0
)
2
; Ebend
=
1
2
Nb
Kb
(Â − Â0
)
2
ERtors
=
1
2
NRt
Kt
( − 0
)
2
(2)
where ks , kb , kt and r0 , Â0 , 0 are the force constants and
equilibrium values for bond stretching, angle bending, and rigid
torsional dihedral angle, respectively. The EFtors term is the poten-
tial contribution for flexible proper dihedral angle « and can be
represented using a five terms Ryckaert–Bellemans (RB) function
[27].
EFtors
=
S
n=0
Cn(cos(«))
n
(3)
The OPLS-AA FF was selected to be used in this work because
the molecular structure of the ACC enzyme in our previous paper
[1] was built according to this FF protocol and most of the potential
parameters needed were available in this force field. The missing
ones were obtained using geometries and energies results cal-
culated with the ORCA quantum chemistry program [28] at the
Hartree–Fock level using 6-31g* basis set. The same procedure used
for creation of the potential energy curves were obtained according
to Kirschner et al. [29], constrained optimizations were performed
by specifying an internal coordinate to be frozen (as can be seen
as a fixed terms in Tables 2 and 3) while all other degrees of free-
dom were allowed to relax fully. The results were analyzed using
Table 3
Dihedral angles 0
and force constants kt
for the SSA model* reported in Fig. 2.
Dihedral ( ) kt
(kcal mol
−1
) 0
O2 Si O3 Si Fixed 37.84
C7 C8 Si O Fixed 178.600
H8 C8 Si O −0.327 66.290
C8 Si O Si Fixed 154.077
C6 C7 C8 Si 0.051 179.820
C3 C4 C5 N 0.200 117.330
H7 C7 C6 N 0.200 61.920
C5 N C6 H6 0.102 118.180
H5 C5 N C6 0.344 0.620

4. G.S. Oliveira et al. / Journal of Molecular Graphics and Modelling 45 (2013) 128–136 131
the GRACE program [30] and parameters obtained are presented in
Tables 2 and 3 as well.
All the Lennard–Jones parameters needed for the non-bonded
interactions were taken from the OPLSAA force-field. The charges
needed to calculate the Coulombic interactions were obtained
using the RESP method [31] implemented in the NWChem program
[32].
2.4. Molecular dynamics simulations
In Section 2.2 a surface model suitable to represent a func-
tionalized AFM tip surface in molecular dynamics simulations was
presented. The dimeric model of the ACC enzyme proposed by
Franca et al. [1] containing 1655 residues (23,185 atoms) was
initially placed 5 ˚A above this surface using a steered molecular
Fig. 4. Schematic representation of enzyme attachment on the SSA surface with some of the possible active sites orientations. ACC enzyme is represented in cartoon view
using VMD program [23].

5. 132 G.S. Oliveira et al. / Journal of Molecular Graphics and Modelling 45 (2013) 128–136
dynamics (SMD) [33,34] protocol. The enzyme was pushed toward
the surface with a constant velocity of 0.001 ˚A/fs during 50 ps until
the first residue reaches 5 ˚A from the surface. After this, the steering
force was removed and a free equilibration was run for 1 ns. The
SSA-ACC enzyme set was immersed in a box containing 143,198
tip3p water molecules [35] and 30 Na+ contra-ions were added. In
Fig. 3 we present a schematic picture of the system under study: (i)
a silanol surface covered by a surface spacer agent (SSA) layer form-
ing a functionalized AFM surface; (ii) an enzyme adsorbed onto the
SSA and the system was solvated.
Molecular dynamics simulations were performed with the
NAMD 2.7 program [24] using standard periodic boundary condi-
tions and the following protocol: (i) after volume adjustment at
1.0 atm by a Langevin piston [36] the calculations were performed
in the NVT ensemble at 310 K and 1.0 atm during 1 ns, and tem-
perature was controlled using a Langevin thermostat; (ii) a cut-off
distance of 16 ˚A was used and long-range corrections were consid-
ered using the Ewald sum formalism [37]. With the cut-off value
used, numerical stability in the trajectory generation using the
NAMD code was achieved. Raut et al. [38] also reported a similar
procedure in their investigation of peptide–surface interaction; (iii)
after (i) a NPT equilibration step with 20 ns was run; (iv) using the
average volume obtained in step (iii) a NVT ensemble calculation
was performed to obtain a new 30 ns simulation trajectory.
The results were analyzed using VMD [23] and GRACE [30] pro-
grams.
3. Results
Some of the force field parameters needed to run molecular
dynamics was calculated as follows:
3.1. SSA force field parameters
The molecular dynamics calculations were performed using
the OPLS-AA parameters. Some SSA model parameters such as
bond distances, bond angles, dihedrals and the corresponding force
constants needed in Eq. (2) were calculated and are presented in
Tables 2 and 3. Geometric mean combining rules were used to
obtain cross interactions potential parameters. As discussed before
these calculations were performed using the ORCA quantum chem-
istry software at the HF/6-31g*level, as mentioned in Section 2.3.
The standard deviation between calculated energy curves and the
ones obtained using the parameters below was 0.05 kcal mol−1.
3.2. Enzyme orientation on the surface
To a further understanding of the ACC enzyme orientation on
the SSA modeled surface a combination of experimental and the-
oretical analysis is needed. The final structure achieved by the
enzyme in the adsorption process depends basically on its ini-
tial orientation on the SSA surface. Experimentally, it is possible
to increase the probability of a determined enzyme orientation
by controlling environment conditions, such as temperature, pH
and reaction time [2,3,39]. From electrostatic considerations the
protonated NH3 groups of an enzyme interacts strongly with
aldehyde groups ( COH). Therefore, the availability of the posi-
tively charged sites on enzyme surface can be used to control its
interaction with the SSA through aldehyde groups [19]. In a previ-
ous study, upon solving the nonlinear Poisson–Boltzmann equation
using a finite-difference procedure, Franca et al. [1] reported the
electrostatic charge distribution of the ACC enzyme. Therefore,
using these considerations, some initial orientations of the enzyme
upon the SSA can be chosen to set initial geometric conditions
for molecular dynamics simulation. As already discussed by those
authors, the enzyme orientation should be chosen keeping the
Fig. 5. Structural mobility of the ACC enzyme after 30 and 50 ns of MD simulation.
Water molecules were not represented.
active site exposed to interact with substrates molecules from the
bulk.
In Fig. 4, three favorable adsorption positions of the ACC
enzyme on SSA surface are presented. The optimal enzyme ori-
entation requires a large surface contact area to ensure strong
adsorption on but living the active sites unblocked to allow
interactions. From the plugin Volarea [40] implemented in VMD
program, the surface contact area values calculated for the three
analyzed areas were: A(P1) ∼= 2623.47 ˚A2, A(P2) ∼= 5000.25 ˚A2 and
A(P3) ∼= 9825.23 ˚A2, therefore increasing from P1 to P3 showing that
P3 is expect to have an optimal adsorption energy. Nevertheless in
the P3 conditions the ACC active sites are not exposed to the bulk
and the interaction with a given substrate is blocked. As the surface
contact area of P2 is greater than P1 and its adsorption conditions
are more favorable, therefore this structure was chosen to proceed
with molecular dynamics analysis.
Using position 2, the setup of initial coordinate was performed
placing the ACC enzyme 5 ˚A above the SSA. The MD simulation was
performed using Steered Molecular Dynamics [34] protocol. In the
equilibration process, the ACC enzyme was enforced toward to the
surface using facilities implemented in the NAMD configuration
file [24]. In this procedure, a constant acceleration of 0.001 ˚A ps−2
(0.13896 pN) was used until the first amino acid had reached 5 ˚A
distance of the surface. Then, a 20 ns NPT simulation was carried out
to equilibrate the system. After this equilibration procedure a 30 ns

6. G.S. Oliveira et al. / Journal of Molecular Graphics and Modelling 45 (2013) 128–136 133
Table 4
Analysis of the hydrogen bond during 30 ns of simulation runs. H-bonds average
detection was resumed for each percentage of the trajectory.
% 25 50 75 100
Enzyme 13 17 23 22
Arginine 6 7.5 6.5 6
Lysine 6 7 7 7.5
of simulation trajectory was performed for further analysis. The
final orientation and energy equilibration curve behaviors during
the simulation process are shown in Figs. 5 and 6, respectively.
One observes that after 20 ns the total energy is almost constant
in agreement with an NVT ensemble calculation.
Using the 30 ns last simulation trajectory some parameters
were analyzed to obtain an insight in the dynamical behav-
ior of the system. Hydrogen bonding analysis between SSA and
the ACC enzyme were obtained monitoring the interactions of
ARG–NH3
+and LYS–NH3
+ enzymatic groups with the aldehyde
group from SSA linker as function of time and the results obtained
are shown in Table 4. The hydrogen bonding populations were
calculated using geometric rules implemented in VMD program
[23] but using structural considerations proposed by Torshin et al.
[41].
In Table 4 the H-bonds analysis shows an average of 18H-bonds
between ACC enzyme and SSA in the last 30 ns of trajectory, as
Fig. 6. Electrostatic, van der Waals and total energies involved during 50 ns of
MD simulation. After 15 ns the system energies stabilized with average of electro-
static energy = −1306.46 kcal mol−1
, van der Waals = −417.682 kcal mol−1
and total
energy = −1724.15 kcal mol−1
.
expected, the main contributions were from arginine and lysine
side groups due to its positive charge.
Another important parameter to be monitored was the enzyme
integrity as a function of time, in order to prevent enzyme
Fig. 7. Monitored dihedral angles ( and ˚) as a function of time for region 2.

7. 134 G.S. Oliveira et al. / Journal of Molecular Graphics and Modelling 45 (2013) 128–136
Fig. 8. Selected (in red) and « (in black) dihedral angles of the region 2: (a, c and e) represent 10 ns of the trajectory after equilibration, and (b, d and f) the last 10 ns of the
trajectory. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Fig. 9. Ramachandran plotting of the ACC amino acids after 50 ns. Most of the amino acids were not affected by the presence of the SSA surface.
denaturation. To perform this analysis, some selected amino acids
were divided into two regions according to their interactivity with
the SSA (Fig. 7). Region 1 is defined as the one with initial interaction
with the SSA and region 2 is the one achieving relevant interaction
after 20 ns. Therefore, the internal flexibility of the regions 1 and
2 was monitored following the dihedral angles values and «.
One should expect different structural behavior in these regions as
a consequence of enzyme–SSA interaction. Nevertheless the dihe-
dral angles distributions as a function of time are almost the same
for both regions. Fig. 8 represents time progression of the selected
dihedral angles from the region 2 in the last 20 ns of the trajectory.
In the first part (Fig. 8a, c and e) dihedral undergo some changes
due the few H-bonds exhibited. In contrast, the last part, the num-
ber of H-bonds was increased (Fig. 8b, d and f), as a consequence,
almost stable values were observed.
One observes better adsorption conditions as a function of
time in agreement with the increasing of the surface contact
area. An average energy/contact area of −88.9546 kcal mol−1 A−2
for region 1 and −102.802 kcal mol−1 A−2 for region 2 were
obtained, to be compared with −52.4201 kcal mol−1 A−2 and
−89.2358 kcal mol−1 A−2 respectively in the beginning of the sim-
ulation (Fig. 7). The increase of surface contact area is clear but
the ACC enzyme active sites are still available to interact with
molecules from the bulk.
The initial and final enzyme structures were compared using the
Ramachandran plotting presented in Fig. 9. Results obtained from
Ramachandran plotting shown no eligible differences between the
and « dihedral angles from the initial and final structure after
50 ns of MD trajectory. To access the dynamic stability of the immo-
bilized ACC enzyme dimeric form an overall deviation from the
starting structure was computed and a RMSD value of 0.22 ˚A was
found. Therefore, the ACC enzyme was not denatured by the inter-
action with the model surface. Nevertheless, one must be aware
that, according to Gunsteren et al. [42], larger trajectory could be
necessary to fully investigate the stability and unfolding of some
enzymatic systems. Regarding the ACC enzyme dimeric system pre-
sented here, such investigation is under way and will be reported
in the future.
4. Conclusions
This paper reports an atomistic model for a functionalized
surface of the AFM tip. This surface is covered by a cross

8. G.S. Oliveira et al. / Journal of Molecular Graphics and Modelling 45 (2013) 128–136 135
linker molecules to the attachment of enzymes for nanobiosensor
purposes. The atomistic model was called SSA, this term stands
for surface-space agent. The OPLS-AA force field was used but
some parameters needed in the atomistic model were calculated at
HF/6-31g level. Molecular dynamic simulations were employed to
simulate the functionalized AFM tip interacting with ACC enzyme,
modeled by Franca et al. [1] in aqueous solution. According to the
electrostatic potential analysis, three possible enzymatic orienta-
tions of the ACC on the AFM tip were proposed. The final molecular
dynamics calculations revealed that after 50 ns time step of simu-
lation, the achieved enzyme geometry preserved the active sites of
ACC enzyme to interact with molecules from the bulk.
From electrostatic potential analysis the enzyme–AFM tip inter-
actions were assumed to be of major importance in two peptide
regions. Hydrogen bonds were monitored as a function of time and
an average of 18H-bonds was found. Most of the H-bonds were
attribute to arginine and lysine side groups with the aldehyde group
of SSA. It was observed that an enhancement of surface contact
area between enzyme–SSA and the actives sites availability to bulk
molecules was preserved. The possibility of enzyme denaturation
due to interactions with SSA was investigated monitoring some
dihedral angles from contact area. Ramachandran plotting showed
that after 50 ns of MD simulation no significant enzyme structure
modifications were detected.
The results obtained show that model proposed in this paper
leads to a further understanding of the enzyme–surface interac-
tions and immobilization process on AFM tips from an atomistic
point of view. It is also worth to note that with minor modifica-
tions the model can be straightforwardly used to study enzyme
immobilization on silica for chromatography applications [43,44].
Therefore, the interactions of a given substrate with an enzyme
active site can be studied considering all the important features
needed for the atomistic modeling in different detection process,
which is helpful in the development of new ligands screening
methodologies, as already discussed by Comer et al. [45].
Finally, the quantum chemical calculations were performed in
Theoretical Chemistry Laboratory at Federal University of São Car-
los whereas the molecular dynamics calculations were carried out
using the Texas Learning & Computation Center (TLC2) super com-
puter cluster facilities at the University of Houston.
Acknowledgements
This work was supported by CNPq – National Council
for Scientific and Technological Development (CNPq – INCT,
573742/2008-1), FAPESP – São Paulo Research Foundation (FAPESP
– INCT, 2008/57859-5, 2007/05089-9,2010/00463-2, 2010/04599-
6), and CAPES – Higher Education Improvement Coordination
Agency. One of us (Oliveira, G. S.) also acknowledges CAPES by
the award of a scholarship during six months he was visiting the
University of Houston, TX, USA.
References
[1] E.F. Franca, F.L. Leite, R.A. Cunha, O.N. Oliveira Jr., L.C.G. Freitas, Designing
an enzyme-based nanobiosensor using molecular modeling techniques, Phys.
Chem. Chem. Phys. 889 (2011) 8894–8899.
[2] C. Steffens, F.L. Leite, C.C. Bueno, A. Manzoli, P.S.D.P. Herrmann, Atomic force
microscopy as a tool applied to nano/biosensors, Sensors (Basel) 12 (2012)
8278–8300.
[3] A. da Silva, D. Deda, A. da Róz, R. Prado, C. Carvalho, V. Viviani, et al.,
Nanobiosensors based on chemically modified AFM probes: a useful tool for
metsulfuron-methyl detection, Sensors 13 (2013) 1477–1489.
[4] F.L. Leite, C.C. Bueno, A.L. Da Róz, E.C. Ziemath, O.N. Oliveira, Theoretical mod-
els for surface forces and adhesion and their measurement using atomic force
microscopy, Int. J. Mol. Sci. 13 (2012) 12773–12856.
[5] M.R. Forim, M.F. das, G.F. da Silva, Q.B. Cass, J.B. Fernandes, P.C. Vieira, Simul-
taneous quantification of azadirachtin and 3-tigloylazadirachtol in Brazilian
seeds and oil of Azadirachta indica: application to quality control and marketing,
Anal. Methods 2 (2010) 860–869.
[6] T. Fisher, A. Oberhauser, M. Carrion-Vazquez, P. Marszalek, J. Fernandez, The
study of protein mechanics with the atomic force microscope, Trends Biochem.
Sci. 24 (1999) 379–384.
[7] P. Desmeules, M. Grandbois, V.A. Bondarenko, A. Yamazaki, C. Salesse, Measure-
ment of membrane binding between recoverin, a calcium-myristoyl switch
protein, and lipid bilayers by AFM-based force spectroscopy, Biophys. J. 82
(2002) 3343–3350.
[8] A. Chen, V.T. Moy, Cross-linking of cell surface receptors enhances cooperativity
of molecular adhesion, Biophys. J. 78 (2000) 2814–2820.
[9] I. Lee, R.E. Marchant, Force measurements on the molecular interactions
between ligand (RGD) and human platelet ␣IIb␤3 receptor system, Surf. Sci.
491 (2001) 433–443.
[10] H. Li, W.A. Linke, A.F. Oberhauser, M. Carrion-Vazquez, J.G. Kerkvliet, H. Lu,
et al., Reverse engineering of the giant muscle protein titin, Nature 418 (2002)
998–1002.
[11] F.W. Bartels, B. Baumgarth, D. Anselmetti, R. Ros, A. Becker, Specific binding of
the regulatory protein ExpG to promoter regions of the galactoglucan biosyn-
thesis gene cluster of Sinorhizobium meliloti – a combined molecular biology
and force spectroscopy investigation, J. Struct. Biol. 143 (2003) 145–152.
[12] P.Y. Meadows, J.E. Bemis, G.C. Walker, Single-molecule force spectroscopy of
isolated and aggregated fibronectin proteins on negatively charged surfaces in
aqueous liquids, Langmuir 19 (2003) 9566–9572.
[13] J. Kaur, K.V. Singh, A.H. Schmid, G.C. Varshney, C.R. Suri, M. Raje, Atomic
force spectroscopy-based study of antibody pesticide interactions for char-
acterization of immunosensor surface, Biosens. Bioelectron. 20 (2004)
284–293.
[14] F. Kienberger, G. Kada, H. Mueller, P. Hinterdorfer, Single molecule studies of
antibody–antigen interaction strength versus intra-molecular antigen stability,
J. Mol. Biol. 347 (2005) 597–606.
[15] C.-K. Lee, Y.-M. Wang, L.-S. Huang, S. Lin, Atomic force microscopy: deter-
mination of unbinding force, off rate and energy barrier for protein–ligand
interaction, Micron 38 (2007) 446–461.
[16] A. Taninaka, O. Takeuchi, H. Shigekawa, Reconsideration of dynamic force spec-
troscopy analysis of streptavidin–biotin interactions, Int. J. Mol. Sci. 11 (2010)
2134–2151.
[17] W.T. Johnson, Immobilizing Biological Molecules on AFM Probes for MRFM and
TREC Studies, Agilent Technologies, 2008.
[18] T. Schroeder, T. Geisler, V. Walhorn, B. Schnatwinkel, D. Anselmetti, J. Mattay,
Single-molecule force spectroscopy of supramolecular heterodimeric capsules,
Phys. Chem. Chem. Phys. 12 (2010) 10981–10987.
[19] A. Etchegaray, C. de, C. Bueno, O. Teschke, Identification of microcistin LR at
the molecular level using atomic force microscopy, Quim. Nova. 33 (2010)
1843–1848.
[20] B. Bhushan, D.R. Tokachichu, M.T. Keener, S.C. Lee, Morphology and adhesion
of biomolecules on silicon based surfaces, Acta Biomater. 1 (2005) 327–341.
[21] F. Richards, J. Knowles, Glutaraldehyde as a protein cross-linking reagent, J.
Mol. Biol. 37 (1968), 231–&.
[22] D.K. Deda, B.S. Brito, C.C. Bueno, A.C.N. da Silva, G.A. Ribeiro, A.M. Amarante,
et al., The use of functionalized AFM tips as molecular sensors in the detection
of pesticides, Materials Research (São Carlos. Impresso 16 (2013) 683–687.
[23] W. Humphrey, A. Dalke, K. Schulten, VMD. Visual molecular dynamics, J. Mol.
Graph. 14 (1996) 33–38.
[24] L.V. Kalé, M. Bhandarkar, M. Bh, R. Brunner, N. Krawetz, A. Shinozaki, et al.,
NAMD: a case study in multilingual parallel programming, in: Proceedings of
10th International Workshop on Languages and Compilers for Parallel Com-
puting, 1997, pp. 367–381.
[25] A. Aksimentiev, R. Brunner, J. Cohen, J. Comer, E. Cruz-Chu, D. Hardy, et al.,
Computer modeling in biotechnology: a partner in development, Methods Mol.
Biol. 474 (2008) 181–234.
[26] W. Jorgensen, D. Maxwell, J. TiradoRives, Development and testing of the OPLS
all-atom force field on conformational energetics and properties of organic
liquids, J. Am. Chem. Soc. 118 (1996) 11225–11236.
[27] J. Ryckaert, A. Bellemans, Molecular-dynamics of liquid normal-butane near its
boiling-point, Chem. Phys. Lett. 30 (1975) 123–125.
[28] F. Neese, The ORCA program system, Wiley Interdiscip. Rev. Comput. Mol. Sci.
2 (2012) 73–78.
[29] K.N. Kirschner, R.D. Lins, A. Maass, T.A. Soares, A glycam-based force field for
simulations of lipopolysaccharide membranes: parametrization and validation,
J. Chem. Theory Comput. 8 (2012) 4719–4731.
[30] Grace: Developer’s Corner 1998.
[31] F.-Y. Dupradeau, A. Pigache, T. Zaffran, C. Savineau, R. Lelong, N. Grivel, et al.,
The R.E.D. tools: advances in RESP and ESP charge derivation and force field
library building, Phys. Chem. Chem. Phys. 12 (2010) 7821–7839.
[32] M. Valiev, E.J. Bylaska, N. Govind, K. Kowalski, T.P. Straatsma, H.J.J. Van
Dam, et al., NWChem: a comprehensive and scalable open-source solution
for large scale molecular simulations, Comput. Phys. Commun. 181 (2010)
1477–1489.
[33] B. Isralewitz, J. Baudry, J. Gullingsrud, D. Kosztin, K. Schulten, Steered molecular
dynamics investigations of protein function, J. Mol. Graph. Model. 19 (2001)
13–25.
[34] B. Isralewitz, M. Gao, K. Schulten, Steered molecular dynamics and mechanical
functions of proteins, Curr. Opin. Struct. Biol. 11 (2001) 224–230.
[35] Y. Wu, H.L. Tepper, G.A. Voth, Flexible simple point-charge water model with
improved liquid-state properties, J. Chem. Phys. 124 (2006) 24503–24512.

9. 136 G.S. Oliveira et al. / Journal of Molecular Graphics and Modelling 45 (2013) 128–136
[36] G. Martyna, D. Tobias, M. Klein, Constant-pressure molecular-dynamics algo-
rithms, J. Chem. Phys. 101 (1994) 4177–4189.
[37] U. Essmann, L. Perera, M. Berkowitz, T. Darden, H. Lee, L. Pedersen, A smooth
particle mesh Ewald method, J. Chem. Phys. 103 (1995) 8577–8593.
[38] V.P. Raut, M.A. Agashe, S.J. Stuart, R.A. Latour, Molecular dynamics simulations
of peptide–surface interactions, Langmuir 21 (2005) 1629–1639.
[39] A.L. Souza, L.A. Marques, M.N. Eberlin, P.A.P. Nascente, P.S.P. Herrmann
Junior, F.L. Leite, et al., Self-assembled hybrid films of phosphotungstic acid
and aminoalkoxysilanes on SiO2/Si surfaces, Thin Solid Films 520 (2012)
3574–3580.
[40] J. Ribeiro, Volarea, Group of Computational BioChemistry, Faculty of Sciences
of the University of Porto, Porto, Portugal, 2011.
[41] I.Y. Torshin, I.T. Weber, R.W. Harrison, Geometric criteria of hydrogen bonds
in proteins and identification of ‘bifurcated’ hydrogen bonds, Protein Eng. 15
(2002) 359–363.
[42] W.F. Van Gunsteren, P.H. Hünenberger, H. Kovacs, A.E. Mark, C.A. Schiffer,
Investigation of protein unfolding and stability by computer simulation, Philos.
Trans. R. Soc. Lond. B, Biol. Sci. 348 (1995) 49–59.
[43] M.C. de Moraes, R.G. Ducati, A.J. Donato, L.A. Basso, D.S. Santos, C.L. Cardoso,
et al., Capillary bioreactors based on human purine nucleoside phosphorylase:
a new approach for ligands identification and characterization, J. Chromatogr.
A 1232 (2012) 110–115.
[44] R.V.C. Guido, C.L. Cardoso, M.C. de Moraes, A.D. Andricopulo, Q.B. Cass, G.
Oliva, Structural insights into the molecular basis responsible for the effects
of immobilization on the kinetic parameters of glyceraldehyde-3-phosphate
dehydrogenase from Trypanosoma cruzi and human, J. Braz. Chem. Soc. 21
(2010) 1845–1853.
[45] J.R. Comer, D.B. Wells, A. Aksimentiev, Modeling nanopores for sequencing
DNA, Methods Mol. Biol. 749 (2011) 317–358.